Environmental Statistics at UCL

UCL Statistical Science has one of the largest environmetrics research groups in the UK, with many staff members and research students involved in its environmental statistics research theme. On 15th October 2024 the theme members spent some time “out of office”, getting to know each other’s work and planning their future activities.

The day revealed the huge diversity of environmental application areas being studied in the department. These include climate modelling and weather simulation; wind farm modelling; risk assessment from natural hazards including tsunami and floods; ecological sampling; the evolution of species; food systems; and mental health interventions to target eco-anxiety.

The discussions also highlighted some statistical issues that recur frequently across a wide range of application areas. These include the analysis of low-frequency, high-impact events such as large floods; the quantification of uncertainty; and the need to work with data that are not representative of the system of interest, for example because they are derived from opportunistic surveys or because some values are missing for reasons that are related to the quantity being measured. The most extreme example of the latter issue involved an analysis of risk from natural hazards, in which the most important wind speed measurements were missing because all of the weather stations in the region of interest were rendered inoperable by a hurricane!

UCL’s environmental statisticians are also active in promoting the uptake of state-of-the-art statistical methods in other disciplines, as well as outside academia: there are active collaborations with industry, non-governmental organisations and – partly via the UCL Met Office Academic Partnership – with the Met Office. Highlights include the development of methodology underpinning much of the analysis in the 2019 and 2023 State of Nature reports for the UK, as well as a supplementary dataset extending the UK’s official climate projections.

A clear take-home message from the day was that there is plenty of opportunity for statisticians to get more involved with the many exciting challenges arising in environmental applications; and also plenty of appetite for collaboration both within and beyond the department, arising in part from the discovery that so many of these application areas give rise to the same fundamental statistical challenges. Watch this space.